Significance of Advection for the Carrying Capacities of Fjord Populations

MARINE ECOLOGY PROGRESS SERIES Published January 2 Mar. Ecol. Prog. Ser. I

Significance of advection for the carrying capacities of fjord populations

' Department of Marine Biology, University of Bergen, N-5065 Blomsterdalen, Norway Institute of Marine Research, Directorate of Fisheries, PO Box 1870, N-5024 Bergen, Norway

ABSTRACT: Advective rates of nutrients, phytoplankton and zooplankton were estimated for the Masfjord, western Norway, in June and October 1985. The advective contribution to the phytoplankton biomass formation was dearly less than the local growth. Advection of nutrients, even below the photic zone, may have large implications for the new production of the fjord. The highest renewal rate (13.6 % dl)due to advection was obtained for the mesozooplankton in June. While 20 % of this renewal was due to water advection alone, 80 % was due to the combined effect of the current profile and the vertical distribution of the mesozooplankton. Thus transport of mesozooplankton may exceed local growth significantly. The deep basin (494 m) of the fjord serves as a refuge for resident macrozooplankton and rnicronekton populations predating on mesozooplankton. Analysis suggests that such predators may be more sensitive to a change in the advective rate than to a similar change in the prey growth rate. Advection may be an important factor underlying the carrying capacity of mesozooplankton predators in fjords.

INTRODUCTION vary. The cross-sectional area above the sill is therefore an important boundary property, and the ratio between Norwegian fjords are presently being considered as a the cross-sectional area and the total fjord volume may biotope for sea-ranching of different organisms. The indicate the impact of the sill boundary conditions on carrying capacities of a fjord, with respect to such the fjord. This ratio varies considerably from one fjord organisms, have therefore become parameters of con- to another. In the fjord system LindAspollene, western siderable interest. Carrying capacity depends on the Norway, this ratio is of order lo-' (m),while the productivity and structure of the system, and a crucial corresponding ratio of Korsfjorden, western Norway, is question is therefore: To what extent are these factors of order lo4.Extensive investigations on the pelagic influenced by boundary conditions rather than by local communities in these fjords have revealed marked processes within the system? Platt & Conover (1971) differences in biological structure (Matthews & Heim- made an intensive 25h study to estimate the terms dal 1980). For example, it has been concluded that (production, transport and grazing) of the continuity Calms finmarchicus in Korsfjorden is heavily influ- equation of phytoplankton formulated for the Bedford enced by 'advective production' (Matthews & Heimdal basin. They found that 58% of daily phytoplankton 1980),while in LindAspollene the population develop- production was lost through exchange (mainly tidal) ment within a year is more influenced by internal across the sill. Lindahl & Perissinotto (1987) and Lin- processes (Aksnes & Magnesen 1983). Extensive dahl & Hernroth (1988) concluded that non-tidal advec- wintedspring renewal of the deep water, however, tion was of major importance in the regulation of the may seriously affect the 'initial' conditions and give rise mesozooplankton biomass of the Gullmar fjord in to yearly variations in species composition and biomass Sweden. (Lie et al. 1983).This was also pointed out by Lindahl & As the physical scale of fjords varies, the balance Hernroth (1988) in their study of the Gullmar fjord. between internal and external forcing is also likely to Masfjorden has been selected as a site for mass release of codlings. As recommended by Ulltang (1984) * Present address: Dunstaffnage Marine Research Laboratory, the ecological and economical consequences of such PO Box 3, Oban, Argyll PA34 4AD, Scotland release should be evaluated within a restricted area

@ Inter-Research/Printed in F. R. Germany 264 Mar. Ecol. Prog. Ser. 50: 263-274, 1989

before large-scale enhancement of natural populations Table 1. Tidal amplitude is 0.5 to lm, and daily is initiated. The study of the factors underlying the exchange due to tide is about 0.5 %. This exchange is carrying capacity for cod is one of the main ecological not included in the rates presented in our study. These topics within this project. are calculated on the basis of residual currents, which In the present study the potential significance of were calculated by a graphic averaging of the current advection in Masfjorden (Fig. 1) is investigated. Mas- component during the observed tidal cycles. The tidal fjorden has a cross-sectional sill area to volume ratio of influence is also left out in the calculation of renewal 8 x lo6which is intermediate between Lindiispollene rates for the biological compartments of the fjord. and Korsfjorden. Results on the amounts and exchange The total freshwater supply varies between 30 and rates (obtained during 2 short time periods) of 60 m3s1, and amounts to about 0.1 % of total fjord nutrients, chi a, phaeopigments, and zooplankton are volume per day. Most of the freshwater enters at the presented, and possible implications of advection are head of the fjord, where the outlet from a hydroelectric discussed. power plant is located. Sampling was undertaken during 2 periods: 24 to 28 June and 28 October to 1 November 1985. (The STUDY AREA AND METHODS latter period will be referred to as the 'October' sampling.) In these periods, continuous current measure- Masfjorden (Fig. 1) is separated from the larger fjord, ments were made at different depths above the sill, Fensfjorden, by a 75 m deep sill. Length of the fjord is while discrete measurements of nutrients, chla, about 20 km and width averages 1 krn. Other important phaeopigments, mesozooplankton and macrozoo- topographic characteristics are given in Fig. 1 and plankton content were made above the sill and within

Fig. 1. Masfjorden with sampling stations. Sl to S3 are located at the fjord sill Aksnes et al.: Advection in fjords 265

Table 1. Topographic characteristics of the fjord frozen for biomass analysis and another preserved in 4 % formalin for species analysis. Dry weight was Sill depth 75 m determined by drying in a oven at 60Â° to constant Maximum depth 494 m weight, and ash content by burning for 2 h at 480Â°C Surface area 2.85 x lo7 m2 The coefficient of variation (CV) between the plank- Cross-sectional area above the sill 4.45 x lo3 m2 ton pump and the Juday-net biomass samples obtained Total volume 5.36 x lo9 m3 Volume above the sill depth 1.80 x lo9 m3 from the same depth layer (but some hours apart) was on average 22 % (based on 7 comparisons). This CV is comparable to the CV calculated from replicate day- the fjord Three sampling stations (Sl, S2, S3) were time sampling with a Juday-net (Aksnes 1986), and located at the sill, and 4 (A, B, C, D) within the fjord present biomass estimates (Table 3) are based on com- (Fig 1) bined data for the 2 gears. Sampling was earned out from RV 'Hgkon Mosby' Macrozooplankton and mesopelagic fishes were and RV 'Fndtjof Nansen' sampled day and night in the deep basin (Stns B and C) Hydrography and current measurements. During 24 and at the sill (Sl) in October. An Isaacs-Kidd Midwa- to 28 June automatic current measurements (Sensor- ter Trawl (IKMT) of 10m2 aperture (Anon. 1981) was data, SD-1000) were conducted at Stns Sl, S2 and S3 fished open at a towing speed of ca 3 knots. The mesh (Fig. 1) At Stn Sl readings were obtained at 2,7, 15,35, size of the trawl decreased from 25 mm at the front to 43, 60 and 80m depth, and at Stns S2 and S3 at 2 and 1.15mm at the cod-end. Depth was recorded using a 35 m Dunng 28 October to 1 November measurements Benthos time-depth recorder (model 1170-1000). The were carried out at the same depths as above at Stn 1, volume of water filtered was estimated by multiplying but no current measurements were undertaken at Stns the trawled distance by the net aperture. Displacement S2 and S3. Supplementary measurements in the sur- volumes of the fishes and macrozooplankton were face layer (0 to 2m) and at intermediate depths were measured separately for each sample. Subsamples of conducted with a Gytre 16 point current meter (Sensor- the macrozooplankton (1/2 to 1/8) were frozen for bio- data) operated from an anchored boat Estimates of net mass determination. Samples were dried to a constant water transport above the sill during the 2 penods were weight in an oven at 60 OC and then burned for 6 to 8 h calculated from residual currents and the cross-sec- at 480Â° to obtain ash weights. A biomass estimate for tional area above the sill each sample was obtained by multiplying the subsam- In June, temperature, salinity and oxygen were mea- pie AFDW by the displacement volume ratio between sured by a Martek-CTDO supplied with 100 m cable. In the total sample and the subsample, and taking the October, temperature and salinity were measured mean value. The biomass estimates presented for the using a Neil-Brown CTD, while oxygen content was macrozooplankton (Table3) were calculated assuming determined in water samples (Niskm samplers) using a net aperture of 10m2, but the effective aperture the standard Winkler technique during sampling is certainly less. These estimates must Nutrients and chlorophyll samples were obtained by therefore be regarded as minimum values. Niskin water samplers and with a plankton pump (see Estimates of total nutrients and mesozooplankton below). Subsamples (30ml) for nutrient analyses were biomass (Table 2) are based on sampling in 2 strata. For given additions of chloroform, and refrigerated until nutrients the approximate position of the nutricline later analysis. Concentrations of nitrate, phosphate and (20 m in June and 100 m in October) was used to divide silicate were measured using a Chemlab Autoanalyser the fjord into the 2 strata. Estimates of the mesozoo- Subsamples (100ml) for chl a and phaeopigment analy- plankton are based on the strata 0 to 50m and 50m to sis were filtered through 0.45 pm Sartorius membrane bottom. Confidence limits were obtained according to filters The filters were frozen, and later analysed by Cochran (1977, p. 95-96) with the use of Satterthwaites' acetone extraction in accordance with Holm-Hansen et approximation formula for calculating the effective al. (1965) number of degrees of freedom. Mesozooplankton sampling was performed in 2 strata, 0 to 50m and 50 to 494m The upper stratum was defined in accordance with the operational range RESULTS of a plankton pump, used to get high vertical resolution in the 0 to 50m The samples were filtered through a Hydrography and residual water transport 180 pm plankton gauze and each sample represented a volume of about 9m3 A double Juday-net (aperture The water masses of Masfjorden may be classified 0 125 m2, mesh size 180 pm) equipped with a closing as: brackish water (0 to 3 m);intermediate water, found mechanism was used m both strata One sample was between the brackish water and the sill depth (75m); 266 Mar. Ecol. Prog. Ser. 50: 263-274, 1989

and deep water, found below the sill depth. The inter- water was also fairly constant, ranging from 4.7 to mediate water may be divided into coastal water with 4.9 mi 1-I (Fig.3). salinity below 34.5%oand Norwegian Trench water with salinity above 34.5 %o. In June a persistent inflowing residual current was Nutrients and biological parameters observed between 2 and 25 to 30m depth, while an outflowing current was observed below this depth Average concentrations of nitrate, phosphate and (Fig. 2). Current measurements conducted at 1.5m silicate below 20m depth were estimated as 12.1, 1.5, depth (not plotted) also indicated an outflowing and 9.1 pM respectively, which give rise to a large current, with a mean velocity of about 10 cm s1over a nutrient reservoir in the fjord (Table 2). The low nitrate tidal cycle. Typical residual current velocites between 2 concentrations above 20m (average 0.4 u.M\ indicate and 25 m depth were 8 to 10 cm s and 6 to 8 cm sP1 nitrogen-limited growth of phytoplankton in June. below 30m. Mean out- and in-transport above the sill Average concentrations of chl a and phaeopigments was about 1600 Â 100 m3sP1in June. The inflowing in the upper 50 m were estimated as 0.4 and 0.2 mg mP3 water was Norwegian coastal water with salinity below respectively. Average mesozooplankton biomass of the 34.3 %O and temperature 8 to 15'C. The salinity of the entire fjord volume was estimated as 6.1 mg AFDW outflowing water was close to 34.6 %o and the tempera- m3. The phytoplankton and mesozooplankton bio- ture was 7.5 to 8OC. masses may be compared by converting the estimates The currents observed in October were relatively of phytoplankton (0.497 tonne chla) and mesozoo- more influenced by the tidal cycle than in June. The plankton (32.8 tonne AFDW) to carbon. Assuming a residual currents were about 4 crns (Fig. 2). Max- carbon to chl a ratio of 50 and a carbon to AFDW ratio of imum observed velocities at inflowing and outflowing 0.5, it is indicated that the phytoplankton biomass (25 tide were 15 to 20 crns.The residual current was 3- tonne C) and mesozooplankton biomass (16 tonne C) layered with outflow in 0 to 10m, inflow between 10 to are of the same order of magnitude. 35m, and outflow between 35 m and the sill depth About 50 % of the total mesozooplankton biomass (Fig.2). Mean net out- and in-transport above the sill was found in the upper 50m (Ttible3). The day and was about 500 k 100 cmsl, a third of the transport night vertical distributions were Similar, apart from the observed in June. Both outflowing and inflowing water 0 to 2m layer which had a 4-fold increase at night above 50 to 60 m depth were classified as coastal water (Fig.2). The numerically dominant mesozooplankton in with temperature 10 to 12 OC (Fig.3). the integrated 0 to 50 m depth layer were Oithona spp., The salinity and temperature of the basin water was Temora longicornis, Evadne nordmanni, Calanus fin- 34.99 %O and 7.4 OC, and the oxygen content of the deep marchicus and Podon spp.

CURRENT (crnlsl MESOZOOPLANKTON BIOMASS Irng AFDWI~~I -12 -8 -4 0 4 8 12

Fig. 2. Vertical distribution of residual currents (averaged for the periods 24 to 28 Jun and 28 Oct to 1Nov), and mesozooplankton biomass (during day and night) as measured above the sill. Upper profiles: June cruise; lower profiles: October cruise. Positive current values indicate transport into the fjord Aksnes et al.: Advection in fjords 267

OXYGEN iml/ll of phosphate in this layer was highest in June (Table2). 0 1 2 3 4 5 6 7 L I I I I 1 I 1 This supports the above suggestion that the phytoplank- SALINITY i%o) ton production was nitrogen-limited in June. 29 30 31 32 33 34 35 The total amount of chl a in the fjord in October was I estimated to be about 0.151 tonne, which is about 30 O/O of the June estimate (Table 2). Converted to carbon (see above) this amounts to 7.6 tonnes which is, again, of the same order as the mesozooplankton biomass (amounts to 10.3 tonne C). The total October amount of mesozooplankton was 60 % of the June estimate (Table3). In the upper 50 m, however, the estimate was only 10% of the June estimate, and 90 O/O of the biomass was located below 50 m in October. The numerically dominant mesozooplankton in the upper 50m were Oithona spp., Pseudocalanus spp. and Acartia longiremis, while Calmsfinmarchicus, Oithona spp. and Microcalanus pygmaeus dominated the layer below. The difference between the 2 layers was also reflected in the mean biomass of individuals: 3 pg AFDW ind.' in the upper layer, and 33 ug ind.' in the deeper layer. The 'minimum' biomass estimate (see 'Methods') of macrozooplankton obtained with the IKMT was about half the mesozooplankton biomass estimate (Table3). At night (Fig. 4) most tows gave estimates between 1 and 2.5mg AFDW m-3. During the day (Fig. 4) almost no macrozooplankton was collected in the upper 50 m layer (at both StnB and Sl),and the samples consisted largely of gelatinous zooplankton. The dominant macrozooplankton species in the deeper part of the fjord were pelagic shrimps (Serges- tes arcticus and Pasiphaea multidentata), euphausids Fig. 3. Vertical distribution of temperature, salinity and oxy- (mainly Meganyctiphanes norvegica), and chaeto- gen at the deepest station (Stn B) in October 1985 gnaths (Eukrohnia hamata and Sagitta elegans). The mesopelagic fishes Benthosema glaciale and In October the total amount ofnitrate in the uppermost Maurolicus muellen were also prominent in the IKMT 20 m of the fjord was about 10 times higher (31.5tonne N) samples (not included in the present biomass esti- than inJune (3.3tonne N).On the other hand,the amount mates). Measured as displacement volume the fishes

Table 2. Estimates of nutrients, chl a and phaeopigment content (given as tonne) of Masfjorden during the 2 cruises. 95 % confidence intervals are given, n: number of samples underlying the estimates

Depth June October strata (m) n tonne tonne

N (as 315k 15 nitrate) 8868k 518

P (as 675 15 phosphate) 16891k 201

Si (as 497k 155 silicate) 1099 2 k 110.0

Chl a 0 152 k0041

Phaeopigments 0 151 5 0 019 268 Mar. Ecol. Prog. Ser. 50: 263-274, 1989

Table 3. Tonnes of mesozooplankton and macrozooplankton biomass in Masfjorden during the 2 cruises. 95 % confidence intervals are indicated, n: number of samples underlying the extimates. Macrozooplankton estimates are based on night samples

Depth Mesozooplankton Macrozooplankton strata (m) June October October n tonne AFDW n tonne AFDW n tonne AFDW

0-50 ? 16.9 L 2.0 14 1.6 + 0.3 6 1.9 Â 1.1 50-494 5 15.9 k 3.9 11 19.1 Â 4.9 4 6.3 Â 3.7 0494 12 32.8 + 11.0 25 20.6 + 11.0 10 8.1 k 2.4

accounted for 37 % of the total volume of the night In June the highest mesozooplankton concentrations samples obtained at StnB. at the sill were associated, both day and night, with the inflowing water between 2 and 30m depth, while the highest concentrations were found in the outflowing Advection of nutrients and biomass layer between 0 and 10m depth in October (Fig. 2). Mesozooplankton exchange rates were much higher in

Daily exchange rates of water were about 2.6 O/O and June than in October (in-transport of 4.47 compared to 1.0 % of the total fjord volume in June and October out-transport of 0.13tonne AFDW d"'; Table 5).This was respectively (Table4). In June the out-transport of due to the lower and deeper biomass (Table3), and to the nutrients was similar to that of water, ranging from weaker residual currents observed in October (Fig. 2). 2.2 to 2.6 % dl,but lower in October, ranging from The influence of advection on the macrozooplankton 0.1 to 0.3 '10 dl.The biological compartments - chla, biomass in October is estimated to be similar to that for phaeopigments and mesozooplankton -were relatively the mesozooplankton (Table 4). It should be noted, more influenced by advection with in-transport rates of however, that the sill IKMT-samples were clearly 6.4, 12.6 and 13.6% dC1 in June, and out-transport dominated by Meganyctiphanes norvegica both in values of 3.1, 4.1 and 0.6 YO dC1 in October (Table4). number and biomass, indicating that this species was the main adverted component in the macrozooplankton/micronekton group. Mesopelagic fishes accounted for 18 % of the displacement volume in samples obtained above 75m at Stn B. At the sill station (Stn S2) this group accounted for 3 %, and only small individuals of Maurolicus muelleri were present. No Benthosema glaciate was obtained at this station. At StnB, however, the biomass of B. glaciate was 8 times higher than the biomass of M, muelleri.

DISCUSSION

Exchange processes

In western Norway the tide may be the most important exchange factor in fjords with a shallow sill. But as the sill becomes deeper the role of other exchange processes increases. Masfjorden is connected with the coastal water through Fensfjorden (Fig. 1). Hydro- graphical observations from the period 1976 to 1980 (unpubl.) have shown that changes in the coastal current rapidly propagate into Fensfjorden and Mas- fjorden. Upwelling of deep water along the Norwegian coast develops 2 to 5d after the onset of northerly Fig. 4. Vertical distribution of macrozooplankton biomass at the deepest station (Stn B) during day (solid line) and night winds. The change in the pressure field generated by (broken line) during the October cruise the upwelling seems to flush out the upper layer on the Aksnes et al.: Advection in fjords 269

Table 4. Daily exchange rates (expressed as % of content inside the fjord) of different materials at the sill. Transport into the fjord is given in the column designated 'In', while outward transport is designated 'Out', A negative 'Net' transport value indicates that net transport is directed out of the fjord

Depth June October strata (m) In Out Net In Out Net

Water 0- 20 20.5 0.9 19.6 3.6 4.5 1.2 0- 50 10.4 7.0 3.4 2.9 3.9 - 1.0 0-494 2.6 2.6 < 0.1 0.7 1.0 - 0.3 N (as nitrate) 0- 20 0.0 0.0 0.0 3.4 2.4 1.0 0-494 0.4 2.6 - 2.2 0.2 0.3 - 0.1 P (as phosphate) 0- 20 15.5 1.1 14.4 3.7 2.4 1.3 0-494 1.0 2.2 - 1.2 0.1 0.1 <0.1 Si (as silicate) 0- 20 4.0 0.5 3.5 2.4 2.9 0.5 0-494 <0.1 2.4 - 2.4 0.1 0.1 <0.1 Chl a 0- 50 6.4 4.6 1.2 1.4 3.1 - 1.7

Phaeopigment 0- 50 12.6 3.6 9.0 2.6 4.1 - 1.5 Mesozooplankton 0- 50 26.2 3.5 22.7 4.3 8.2 - 3.9 0-494 13.6 2.8 10.8 0.3 0.6 - 0.3 Macrozooplankton 0-494 0.3 0.6 - 0.3 same time scale (Saetre et al. 1988). The processes western Norwegian fjords, and an out-transport in the resulting in downwelling near the coast outside Fens- lower part. Our June situation, with inflowing water fjorden are more complicated. In addition to southerly above 25 to 30m depth, was most probably associated winds, sudden outflows of water from the Skagerrak with coastal downwelling. This is supported by the (southeast of Norway) may result in downwelling (Aure wind data given in Fig. 5. After a period with northerly & Sastre 1981). The effect of such outflows may last for winds in early May, southerly winds became more 10 to 20 d. Between periods of marked coastal up- and prominent during late May and June. downwelling, local processes (wind, tide and land run- In late October no marked transport in the intermedi- off) become more important for the current regime of ate layer was observed. Coastal water was present the fjord. above 50 to 60m depth, and the situation may be Coastal downwelling leads to in-transport of coastal characterized as a calm period after in-transport of water in the upper part of the intermediate layer of coastal water in the upper layer. The dominating south-

Table 5. Daily transport rates at the sill. Water is expressed as 10m3 dl,other parameters as tonne d-'

Depth June October strata (m) In Out Net In Out Net

Water 0-20 112 5 107 20 13 7 0-50 133 90 43 38 50 -12 0-75 142 138 -4 38 51 -13 N (as nitrate) 0-20 0.0 0.0 0.0 1.06 0.77 0.29 0-75 2.94 20.87 -18.07 2.10 2.66 -0.56 P (as phosphate) 0-20 1.24 0.08 1.16 0.25 0.16 0.09 0-75 2.32 4.98 -2.66 0.48 0.87 -0.39 Si (as silicate) 0-20 1.00 0.12 0.88 1.17 1.46 -0.29 0-75 5.46 30.89 -25.43 2.21 3.63 -1.42 Chl a 0-50 0.03 0.02 0.01 0.002 0.005 -0.003 Phaeopigment 0-50 0.04 0.01 0.03 0.004 0.006 -0,002 Mesozooplankton 0-50 4.43 0.59 3.84 0.07 0.13 -0.06 0-75 4.47 0.91 3.56 0.07 0.13 -0.06 Macrozooplankton 0-75 0.03 0.05 -0.02 270 Mar. Ecol. Prog. Ser. 50: 263-274, 1989

Fig. 5. Daily wind observations in Bergen during 1985. Positive y-values indicate southerly winds, neqativc values indicate 'J'F M'A'M'J'J'A'S'O'N'D' northerly winds

orly winds during late September and October [Fig. 5) sill. In contrast to the above species, the euphausid are probably responsible for the presence of the thick Meganyctiphanes norvegica seems to be heavily influ- layer of coastal water, enced by advection at night. According to the above presentation, a high frequency of alterations between southerly and northerly winds leads to an increased influence of advection. As Nutrients, chlorophyll and mesozooplankton demonstrated in Fig,5 this frequency may be quite high. During 1985 the persistence of northerly wind If the different nutrients and organic materials had events was on average lower than the persistence of been distributed homogeneously in the fjord (and southerly winds. With a too low persistence of northerly assuming passive advection) the influence on them of winds the role of advection may diminish. Both the advection would have been the same as for water, i.e. frequency and the time scales of northerly wind events an exchange rate of 2.6 "lo dlin June and 0.7 to 1.0 % vary from one year to another, and these characteristics d1in October (Table 4). The actual exchange rates are probably the main forcing for the annual renewal of (using the highest of the 2 values 'In' and 'Out' in Table the intermediate layer of fjords located on the west 5) for nutrients are equal to or lower than these figures coast of Norway. We suggest that a strong wind forcing on both sampling occasions. This is to be expected may propagate effectively, through advection of the since the highest nutrient concentrations of the fjord intermediate layer, into the biological compartments of are located below the advective layer. The daily fresh- the fjord. This will be discussed later, but first we will water run-off amounts to only 0.1 % of the fjord volume. pay some attention to the deep water of the fjord. It has a significant influence, however, on the upper The deep water of Masfjorden (below the sill depth) 2m. Our June measurements (not tabulated) indicate is generally not influenced by extensive transport pro- that the freshwater run-off represents a direct input to cesses. One total exchange was observed during the the photic zone of about 0.4 tonne Ndl. period 1976 to 1980 (unpubl.). The minimum oxygen Contrary to the nutrients, chla and phaeopiyments concentration of the deep water was 3.7 milin this were confined to the advective layer of the fjord. Dur- period. ing both cruises the exchange rates were 3.1 to 4.8 times higher than for water, ranging from 3.1 to 12.6% d1(using the highest of 'In' and 'Out' in Table 4). Such Macrozooplankton and mesopelagic fishes enhanced exchange was even more pronounced for the in the deep water June mesozooplankton with an in-transport of 13.6% d ', which is 5.2 times the exchange rate of water. This The deep water represents a potential refuge for means that 20% of the mesozooplankton exchange pelagic organisms local to the fjord, and offers a was due to water advection alone, while 80 % was due homogeneous environment with respect to tempera- to the combined effect of the current profile and the ture and salinity (Fig. 3). This seems to he utilized by vertical distribution of the mesozooplankton. the pelagic shrimps Pasiphea multidentata, Serqestes To what extent the planktonic part of a fjord system arctica and the mesopelagic fishes Benthosema is controlled by internal biological processes rather qlaciale and Maurolic~~smuelleii. Although these than advective processes depends on the physical scale species perform die1 vertical migrations into the advec- of the fjord versus the time scale of these processes. As live intermediate layer of the fjord, it is likely that they a general simplification we may write: are able to sustriin their horizontal distribution within the fjord (Kaartvedt et al. 1988). Excluding small num- bers of Maurolicus niuelleri, none of these species were present in either day or night samples collected at the where B = biomass concentration within the system Aksnes et al.: Advection in fjords 27 1

(mg m3);t = time (s);r = local instantaneous growth results where 90 % of the mesozooplankton (Table3) rate of B (s);v = mean absolute current above the sill was distributed below 50 m. This distribution resulted in (msl); BB = biomass concentration in incoming an exchange value (0.6 % dl)that was lower than the current (mgm3);A = cross-sectional area above the water exchange value (0.7 to 1.0 % dl).A deep localiza- sill (m2);V = fjord volume (m3), tion of the mesozooplankton during late autumn and As B approaches Bn the net advective effect becomes winter is the normal situation in Norwegian fjords. zero. This does not mean, however, that the advective From Eqs. (1) and (2) we see that the value of R (ratio effect has ceased, since the biomass renewal within the between the cross-sectional area above the sill and the system still may be dominated by advection rather than fjord volume) gives the order of magnitude of the local growth. The growth rate (r) and the advective rate advective influence in a particular system. This ratio ((3 = 0.5vR) have the same dimension (sl),and the ratio varies considerably from one fjord to another (see r/(3 decides which of the 2 processes dominates the 'Introduction'),and it may serve as an index indicating biomass formation within the system. If r/B > 1,growth is the potential advective influence on a system. As the dominating process while r/B < 1 indicates advective shown above, the product of R and the typical current dominance. The importance of advection relative to the velocity (v) across the boundary surface (A) is a quan- growth of phytoplankton and zooplankton is indicated tity with a dimension (s)equal to the dimension of in Fig. 6. The much lower growth rate of zooplankton biological processes rates. A direct comparison of these compared with phytoplankton implies that transport rates may therefore be performed. influences primarily the zooplankton biomass (Fig. 6). The highest exchange rate of chl a was 6.4 % d.In Phytoplankton is, however, constrained to the upper, June, a phytoplankton doubling time of 1 d is not photic zone where transport processes are most promi- unrealistically high. This corresponds to a 100 % nent. Zooplankton may utilize the entire water column increase each day, and local growth was probably the and the advective influence may thereby diminish. A dominating process of the phytoplankton biomass quantitative effect of this is demonstrated in Fig. 6. We renewal in the fjord. see that the zooplankton confined to the advective layer The primary production seems to be strongly limited (dotted line) is 3 times more influenced by advection by the availability of nitrate in June, which is probably than zooplankton distributed in the entire fjord volume the normal situation during summer. Assuming a N : chl (solid line). Similarly, vertical migrations (die1and sea- ratio of 8 and an average phytoplankton doubling time sonal) may also reduce the influence of advection for of 1 d, the estimated phytoplankton stock of 0.497 populations depending on the food availability in the tonne chl a within the fjord (Table 2) may utilize about advective layer. This is demonstrated by the October 4tonne nitrate dl.The 0.4 tonne nitrate supplied by the freshwater run-off accounts for only 10 % of this potential nutrient uptake. Factors influencing the transport of new nutrients into the photic zone from below are therefore of major importance for the new Zooplankton (0-70rn) ...... ' production within the fjord...... ' While no advection of nitrate was measured in the ...... upper 20m in June (Table5), an amount of 18 tonne nitrate d was exported below this depth (Table5). Although most of this nitrate was exported below the photic zone, such losses may affect the future vertical transport of nitrate from non-photic to photic zone. Hence, advective nutrient loss in the layer close to the photic zone may lead to reduced new production within the fjord. The loss of 18 tonne nitrate d-I exceeds the assumed daily growth by a factor of 4, and advective nutrient exchange (even below the euphotic CURRENT (cm s-') zone) may therefore be of greater importance to the Fig. 6. Scale analysis on the role of advection relative to primary production than the exchange of the phyto- production. Y-values above 1 indicate that advection plankton itself. dominates over production in the formation of plankton bio- The above nutrient loss may be turned into nutrient mass in the fjord. The x-axis represents the mean absolute supply in the case of northerly winds. In this situation current above the sill. Plankton is assumed to be distributed homogeneously throughout the depth interval indicated. the lower part of the intermediate layer may be Growth rate of zooplankton was assumed to be 5% d-I; enriched by nutrients, and thereby give rise to an doubling time of phytoplankton was assumed to be 1 d increased new production. 272 Mar. Ecol. Prog. Ser. 50: 263-274, 1989

Earlier observations at the head of Masfjorden Masfjorden. The diet of B. glaciale and M. muelleri is mesozooplankton (Gjesaster 1973, 1981). Daily measurements of the zooplankton biomass in Coastal cod has long been regarded as geo- the intake water of the aquaculture station in Matre graphically stationary (Dahl 1906), and release experi- (Fig. 7) were made in the years 1976 to 1978 (Anon. ments with artificially propagated cod (Svgsand 1985) 1980). The seawater intake of the station is located at support this. The nearshore living Gobiusculus flaves- 10m depth at the head of the fjord (inwards of Stn D; cens and the mesopelagic shrimps and fishes are also Fig. I), and the zooplankton measurements were believed to be part of a local fjord community, at least obtained by filtering (500 urn) the intake water. The when compared to the mesozooplankton. We therefore samples were dominated by Calanus finmarchicus. A think that the relation between the stationary mesozoo- physical/biological aggregation mechanism rather plankton predators and the adverted prey is an impor- than local production is probably the explanation for tant feature of fjord ecology. the very high biomass observations (Fig. 7). The bio- A simplified relationship between a stationary preda- mass build up corresponds to an average increase as tor (C) in the fjord, and a zooplankton prey (B) influ-

high as 10 to 20 O/O dl. Such an increase may be enced by both local growth and transport may be explained by advective rates similar to those observed expressed: in June. At this time, the net in-transport of mesozooplankton at the sill was estimated to be 22.7 % d in the upper 50m (Table 4). The marked decrease in biomass in May and June (Fig. 7) may be due to advec- where B = prey concentration within the system; BB = tion, mortality or migration. As a result of downward prey concentration at the boundary; C= predator con- migration, Calanus finmarchicus is known to disappear centration within the system; r= instantaneous growth from the upper water column in Norwegian fjords dur- rate of B when predator is absent; K= carrying capa- ing late spring (Aksnes & Magnesen 1983). city (equilibrium biomass attained by prey when predator and advection are absent); a = predation constant; 6 = instantaneous exchange rate of the system Mesozooplankton advection and higher trophic levels due to advection, corresponds to O.5vR in Eq. (1); e= conversion efficiency; d= death rate of predator (C). Mesozooplankton is the main diet of cod fry less than If the net advective rate, b = 0 (no advection of prey), 5 cm, while the cod between 5 and 20 cm feed predo- the above equations correspond to the Lotka-Volterra minantly on Gobiusculus flavescens in Masfjorden equations. (Salvanes 1986). Investigations on this species are The equilibrium biomasses of prey (B') and predator currently undertaken and the diet is reported to be ((7) are: mesozooplankton (Fossd pers. comm.). In fjords of northern Norway Maurolicus muelleri, Benthosema qlaciale and Pasiphaea sp. are reported to be prey items of cod (Santos & Falk-Petersen 1988), and these According to this equilibrium solution, the prey equi- species are also possible prey items for the cod in librium biomass (B') is not influenced by the exchange

Fig. 7. Zooplankton content (5d running mean) of the intake water of the Matre aquaculture station located at the head of the fjord (inwards of Stn D) (redrawn from Anon. 1980) Aksnes et al.: Advection in fjords 273

rate (13) or the boundary biomass (B-a}. The predator levels. It is shown that a stationary predator relying on biomass (C'),however, responds linearly to a change in the availability of mesozooplankton in Masfjorden may both the exchange rate and the boundary biomass, be strongly influenced by mesozooplankton advection. We may rewrite Eq. (6): Young cod belong to this group of predators. Both the actual level of biomass and the stability of the fjord C' = l/a (r[l - B1/K\+ b), populations may depend on the strength and variability where b may be termed the net advective rate of the of the transport processes. In an advective system the prey, which corresponds to the net exchange rate in resident mesozooplankton predators are not a part of a Table 4 (expressed as a finite percentage in the Table). usual predator prey relationship as described by the We see that a change in the growth rate (r)of the prey classical Lotka-Volterra equations. In the advective also affects the predator equilibrium biomass (C1) system the predator cannot exterminate their prey and linearly, but the effect is suppressed by the factor in that sense the system may be stabilized. On the other 1 - B'/K. This means that the predator biomass is more hand, variability in the advection itself may propagate sensitive (depending on how close the prey biomass is into the system. Furthermore, most predators have to its carrying capacity) to a change in the net advective pelagic larvae that are transported via currents, and rate (b) than to a change in the growth rate of the prey recruitment may therefore be influenced positively or (r).This, of course, relies on the assumption that the negatively by advection. Future research concentrating advective biomass is equally available to the predator on the impact of advection on an annual basis (the time as the biomass originating from local production. The scale of the predators' life cycle) is necessary in order to estimates of b for the mesozooplankton were 0.10 d-I obtain a thorough understanding of the coupling (corresponds to 10.8 % in Table4) in June. Assuming a between advection, stability and carrying capacities of growth rate for the entire population of 0.05d-I and predators in fjords. that the 'equilibrium' biomass was half the carrying capacity, the 'local' contribution to the zooplankton Acknowledgements. We thank A. Aadnesen for field and predator would have been 0.025 while the advective laboratory assistance, and U. Lie, T. Heisaster, J. H. FossA and J. contribution would have been 4 times this figure (0.10). Giske for stimulating discussions and commenting on the manuscript. We also acknowledge the meteorological data It must be noted that if the upper 0 to 50 m is considered given by the Norwegian Meteorological Institute, and the separately the significance of advection is even more financial assistance from Fiskeridirektoratets Effek- pronounced (22.7 % dC1; Table4). The estimate of b in tiviseringsmidler and from the Newfoundland Institute for Cold October was considerably less and had a negative sign Ocean Science and the Canadian-Scandinavian Foundation. (-0.003). At this time, however, the growth rate of the mesozooplankton was probably also suppressed. LITERATURE CITED

Aksnes, D. L. (1986).Resource allocation in the study of copepod CONCLUSIONS population dynamics. Dr scient. thesis, Univ. Bergen Aksnes, D. L., Magnesen, T. (1983). Distribution, develop- The growth rate of phytoplankton (in the productive ment, and production of Calanus finmarchicus (Gunnerus) season) is probably high enough to dominate the bio- in Lindaspollene, western Norway, 1979. Sarsia 68: mass renewal in the Masfjord. This seems to be true 195-208 Anon. (1980). Havmilje og Havundersekelser i de norske even at times with high advection across the sill bound- fiskeriomrader. Fisken og Havet 1980 (Sasrnummer 1): ary. During summer, supply of nutrients from below 1-61 may be the main nutrient source for new production. Anon. (1981).Manual for the international young fish surveys Advection of nutrients in the intermediate layer below in the North Sea, Skagerak and Kattegat. Coun. Meet. int. Coun. Explor. Sea C.M.-ICES/H: 9, p. A.0.-B.3 the photic zone may be high enough to severely influ- Aure, J., Sastre, R. (1981). Wind effects on the Skagerak ence the future vertical transport of nutrients into the outflow. In: Seetre, R., Mork, M. (eds.) The Norwegian photic zone. coastal current. University of Bergen, p. 263-293 The wind regime along the Norwegian coast is prob- Cochran, W. G. (1977). Sampling techniques, 3rd edn. John ably the main forcing for the renewal of the intermedi- Wiley & Sons Inc., New York Dahl, K. (1906). Undersegelser over nytten av torskeudkleek- ate layer of the Masfjord. Variability in this wind ning i estlandske fjorde. Betenkning. Arsberetn. Norg. regime may create variability in the new production of Fisk. 1: 3-97 the fjord. This variability may further propagate into a Gjesaster, J. (1973). The food of the myctophid fish Ben- variability in the carrying capacity for the higher thosema glaciale (Reinhardt),from western Norway. Sarsia 52: 53-58 trophic levels. Gjeseeter, J. (1981). Life history and ecology of Maurolicus The advection of mesozooplankton results in a more muelleri (Gonostomatidae) in Norwegian waters. FiskDir. direct influence on the carrying capacity of higher Skr. (Ser. Havunders.) 17: 109-131 274 Mar. Ecol. Prog. Ser. 50: 263-274, 1989

Holm-Hansen, O., Lorenzen, C. J., Holmes, R. W., Strickland, Platt, T., Conover, R. J. (1971).Variability and its effect on the J. D. H. (1965). Fluorometric determination of chlorophyll. 24 h chlorophyll budget of a small basin. Mar. Biol. 10: 52-65 J. Cons. int. Explor. Mer 30: 3-15 Salvanes, A. G. V. (1986). Preliminary report from a compara- Kaartvedt, S., Aksnes, D. L., Aadnesen, A. (1988). Winter tive study of the diet of four gadoid fishes in a fjord of distribution of macroplankton and micronekton in Mas- western Norway. Coun. Meet. int. Coun. Explor. Sea C.M.- fjorden, western Norway. Mar. Ecol. Prog. Ser. 45: 45-55 ICESIG: 71: 1-18 Lie, U., Magnesen, T., Tunberg, B., Aksnes, D. (1983). Prelimi- Santos, J. DOS,Falk-Petersen, S. (1988). Feeding ecology of nary studies on the vertical distribution of size-fractions in cod (Gadus morhua L.) in Balsfjord and Ullsfjord, North the zooplankton community in Lindaspollene, western Norway, 1982-83. J. Cons. int. Explor. Mer (in press) Norway. Sarsia 68: 65-80 Sastre, R., Aure, J.,Ljeen, R. (1988).Wind effects on the lateral Lindahl, O., Hernroth, L. (1988). Long-scale and long-term extension of the Norwegian Coastal water. Cont. Shelf variations in the zooplankton community of the Gullmar Res. 9: 239-253 fjord, Sweden, in relation to advective processes. Mar. SvSsand, T. (1985). Preliminary results form tagging and Ecol. Prog. Ser. 43: 161-171 release of artificially reared 0-group coastal cod (Gadus Lindahl, O., Perissinotto, R. (1987). Short-term variations in morhua L.) in western Norway. Coun. Meet. int. Coun. the zooplankton community related to water exchange Explor, Sea C.M.-ICES/F: 9: 1-18 processes in the Gullmar fjord, Sweden. J. Plankton Res. 9: Ulltang, 0. (1984).The management of cod stocks with special 1113-1132 reference to growth and recruitment overfishing and the Matthews, J. B. L., Heimdal, B. R. (1980). Pelagic productivity question whether artificial propagation can help to solve and food chains in fjord systems. In: Freeland, H. J., management problems. In: Dahl, E., Danielssen, D. S., Farmer, D. M., Levings, C. D. (eds.) Fjord oceanography. Moksness, E., Solemdal, P. (eds.) The propagation of Cod Plenum Publ. Corp., New York, p. 377-398 Gadus morhua L. Fledevigen rapportser. 1, p. 795-817

This article was submitted to the editor; it was accepted for printing on September 28, 1988